Environmental Applications of Computational Fluid Dynamics - Environmental Modelling: Finding Simplicity in Complexity - page 97

being extended to consider flow around and through bed
features such as pebbles (Hardy et al ., 2005). Although
these results require substantial computing resources and
may be some way off practical application, they offer a
means of understanding the fundamental processes of
turbulence in river channels.

these methods were applicable, the concepts of different
layers of the flowwere derived - in particular the idea that
there was a relatively thin region close to the surface of
the topography where the effects of the topography and of
turbulence were important, and a region above that that
could be regarded as inviscid to a first approximation.
This concept was central to the work of Jackson and
Hunt (1975) who formalized it in their linear analysis of
turbulent flow over low hills, defining an inner layer close
to the hill and an outer layer above that. This theory led to
quantifiable predictions of wind speed-up and the change
of turbulent structure over a variety of topographical
shapes. When tested against actual full-scale experimental
data obtained from controlled measurements over simple
topography (such as the Black Mountain, Canberra,
Australia, and Askervein Hill, South Uist, Scotland) this
theory was shown to perform more than adequately. The
theory continued to be developed over the next decade
or so (Hunt et al ., 1988; Belcher, 1990; Belcher and Hunt,
1998) and forms the basis of a number of calculation
methods that are currently in use for the prediction of
wind fields and dispersion over arbitrary topography.
An example of this is provided in the work of Inglis
et al . (1995) for the prediction of wind conditions over
Kintyre (which was carried out to assess the impact of
wind on forests). Typical comparisons of the predictions
with theory are shown in Figure 6.4.
Wood (2000) has expressed scepticism that standard
closure schemes (Reynolds-averaged Navier-Stokes or
RANS: see Equations 6.3 to 6.6) can adequately repre-
sent flow over large-scale topography. He advocates the
continued development of LES models in this field, and
this seems to be generally accepted as the way forward for
the use of CFD techniques for such problems. It has the
particular advantage of being able to predict flow fluctua-
tions, extreme values etc, which are often the parameters
required in practical situations.

6.3.1.3 Small-scale river works

Three-dimensional CFD software can also be used for
analysing flows at small-scale river works such as sluice
gates, weirs, outfalls and fish passes (Rodrıguez et al .,
2001). This work is neither straightforward nor rapid,
but with further verification (Hargreaves et al ., 2007) is
becoming an acceptable alternative to scale models. In
these cases CFD has the advantage over physical models
of being able to work at full scale.

6.3.2 Atmosphericapplications

In this section we consider the application of CFD tech-
niques to two scales of atmospheric flows - large scales,
greater than 1 km, and small scales, those flows on
an (intra-) urban scale, between 10 to 100m. For the
small-scale flows, the discussion will be referred to as
computational wind engineering (CWE). The applica-
tion of CFD techniques to the prediction of large-scale
atmospheric circulations is not discussed. For details of
such applications, see Mason (1994).

6.3.2.1 Wind over large-scale topography

The calculation of wind speeds and other atmospheric
variables, such as temperature and moisture, over large-
scale topography such as hills, is of practical importance
for assessing local wind climate for planning purposes,
the positioning and optimization of wind farms and
the dispersion of atmospheric pollutants. It also can be
viewed as ameans bywhich larger scale numerical weather
predictions can be used as boundary conditions for small-
scale, embedded models. An excellent historical review of
the subject has been presented by Wood (2000) and the
following few paragraphs draw heavily on this work.
Work in this field began in the 1940s and 1950s with
inviscid flow calculations over a range of large scale
topography. The work of Scorer (1949) on trapped lee
waves stands out as being of considerable importance.
Whilst these calculations were of fundamental impor-
tance they could not, of course, take into account the real
surface boundary condition, or the turbulent nature of
the boundary layer. To assess the flow regions in which

6.3.2.2 Computational wind engineering
The use of CFD techniques at the urban scale is becoming
increasingly common both within the research com-
munity and within an industrial context. The proceed-
ings of the five recent conferences on computational
wind engineering (CWE) - Tokyo, 1992; Colorado, 1996;
Birmingham, 2000; Yokohama, 2006 and Chapel Hill,
2010 - indicate the considerable effort that is being put
into developing this field. The use of CWE to predict wind
effects on pedestrians, pollution dispersion and, to a lesser
extent, wind loads on buildings and other structures is